ProbeSelection Probabilistic methods for identifying signatures of selection on evolutionary trees: from data to theory and back

Selection is one of the main forces that shapes the genetic diversity observed in an evolving population, but characterizing its strength and patterns is still challenging. B-cell affinity maturation is an example of accelerated evolution under selection, which allows us to study these processes on shorter timescales than macroevolution. Here we will develop new methods to learn the properties of B cell selection from repertoire sequencing data, overcoming two issues: current methods do not allow us to go beyond neutrality (absence of selection) to infer evolutionary processes; we lack good summary statistics estimators to go beyond rejecting neutrality. We will build mathematical models of proliferating cell populations undergoing non neutral mutations, and use them to characterize patterns of selection in trees. Informed by those models we will develop inference schemes to learn the parameters of selection from the data. The outcome will be tools to characterize selection, and a basis for identifying responding lineages in clinical settings.